GitLab

Use case: PTS test case for cross transport link key
derivation

Steps to reproduce:
1. Start Test case GAP TP/LEP/DM/BV-15 in PTSv6.0

Failure: PTS test case TP/LEP/DM/BV-15 fails as the link
key is not stored when the key is derived from other
transport key. So it will again go for pairing even though
it has link key but not stored.

Root cause: Link key is stored only when bond type is
presistent, but it sets the bond type to persistent only
when it receives SSP request but not link key is derived.

Fix: So making the solution generic to store always when the
link key type is authenticated combination key with P-256.

Bug: 22486860
Change-Id: I3eb201262f72c86a9438351cad14a6698065fa90

As per feedback from manufacturer.

Bug: 21817410
Change-Id: I4b5b3d8d99f9299754221968a5e1ca5f21cb9478

Bug: 21817410
Change-Id: I5c5f6da9e8326e4c9424907337331b39fb67b444

Change-Id: I7177aaa0e2d0c3b0fa79ae655c0171f0703c54f3

As a side-effect to increasing BTM_BLE_CONN_INT_MIN, valid connection
parameter updates can be rejected if the requested minimum is < 11.25ms,
but the requested maximum is >= 11.25ms.

This patch introduces a new parameter "BTM_BLE_CONN_INT_LIMIT" and
resets BTM_BLE_CONN_INT_MIN to the HCI spec defined minimum.

Bug: 22245224
Change-Id: I6a2e4d18488b63570f7d0070413d00f16a7fdba1

Assign appropriately the p_rsp_data pointer to point in the
buffer with the response.

Bug: 22307858
Change-Id: I2f77f290ff2c62c09b3be18ff047167a48c7c9d4

This change allows the host peripheral which initiates the pairing
to reject the pairing in case the remote does not support
secure LE connections and the host DUT is in secure connections
only mode.

Bug: 22203134
Change-Id: If58f791cb575d6b66c361f58e574b613d5686047

Use case: Reconnection of legacy mouse.

Failure: Not able to reconnect with mouse if the previous
connection rejected was with SC enabled device.

Test Steps:
1. Pair and connect remote devices.
2. Start streaming music over headset and Disconnect HID
   mouse.
3. Inquire and try to pair with any DUT which supports
   secure connection.
4. Cancel pairing and Turn OFF remote headset.
5. Now try reconnection from HID mouse.

Root Cause: When the secure connection is established, it will
have the 3 pages filled, and when the same acl entry is reused
for next mouse reconnection, only one page is getting over
written because legacy HID will have only one page supported
and remaining 2 pages still holds the previous device's data.

Fix: Fixed it by resetting the ACL entry when the ACL is
disconnected so that even though one page is copied and
remaining pages will not contain valid data.

Bug: 22249108
Change-Id: Icd96f22fe0787ef5845ee02940c2b0eacf7a2bbd

This CL prevents functions, bta_ag_hfp_result and bta_hf_client_send_atd, from
the potential strings overflowing.

Bug: 20674686,20677309
Change-Id: Iaef720fc784e020f237feb86e17857bddf57bbfc

A local struct variable is not initialized (memset to 0)
when sending GET_REPORT_EVT to an application.
As a result, field hs_data.rsp_data.p_rpt_data might be used
uninitialized.

Bug: 22352578
Change-Id: I72141117187c392bc0502a17bf71859e67c8da54

This happens in race conditions when we try to connect to a remote
A2DP device from Bluetooth Settings UI and simultaneously there is incoming
connection request from previously connected remote. This fix addresses
the scenario where because of first A2DP connection btif moves to opening/
opened state and then the second connect request is received.
This scenario is handled by gracefully rejecting the second connect request
from btif layer if it is already connected or about to be connected by the
first connect request.

Bug: 22028876
Change-Id: I53be400d1d7c9a84ca9d9a75b6b6f010975cf0cf

Handled the issue of incorrect address type and value
(on Privacy 1.1 chipsets) on reconnection by setting
the correct device type and address value on advertiser end

Original author: Chaojing Sun <cjsun@broadcom.com>

Bug: 22100659
Change-Id: I009888cc9ebd0b631543027d6e34b6c7ecbd0ce8

- Remote device may initiate AVDTP Suspend within 3s of AVRCP pause,
  which sets remote suspend flag. Later on resumption of local playback
  A2DP packets can not be processed further as Remote suspend flag
  remains set. current change overrides remote suspend flag if local
  suspend is triggered later.
- This change covers the usecase where play is pressed from DUT UI
  before local suspend triggers in such scenarios current change takes
  care of resetitng remote suspend flag on AVRCP play state change event.
- Moreover this change will drop A2DP packets when the setparameter
  is called with A2dpSuspended=true and resumes when the mentioned parameter
  is reset.

This change along with corresponding app change helps mitigating
issues reported due to A2Dp, HFP synchronizations and local and remote
device initiated AVDTP suspend handling.

Bug: 22028876
Change-Id:  I3bb334373414170254744d9233cbd2e66219bfa0

Use case: Crash observed during BT IOT testing

1. Try to pair to the remote device.
2. Connect  to the remote device's audio profiles.
3. Remote device has given the wrong UUID's length.This will leads to crash.

Failure: crash observed during profile connection

Root cause: Remote sent invalid UUID length,which is causing crash
in comparison logic.

Fix: Validate input UUID's length before going for actual comparison.

Bug: 19417758
Change-Id: I8216d17e3f6cc22dfbeca4b31972b5b5584a73ea

Bug: 21817410
Change-Id: I55b3753b0ae9f389f1531bb588c4868e96842aeb

Problem: Device shows up in paired list during just work pairing

1. Create a connection from DUT to remote.
2. Remove the link key in remote.
3. Send a file from remote to DUT.
4. Turn OFF BT. Turn ON.

Failure: Device should not move to paired list after BT on/off
as device removed from UI during just work pairing

Root cause: remote device is not removed from btif but
its removed from UI during justworks pairing

Fix: remote device is removed from btif during justworks
pairing to sync up UI and btif paired list

Bug: 19417758
Change-Id: I73de0a94c4f7ada644d258790a9e3e4bf11fc4e5

Previously, only the last snoop log was saved, which was often not
enough if the bluetooth stack was in a crash loop or an app was
resetting it frequently.

Bug: 22202788
Change-Id: I74622ceabe9ce12d5ba03c640c37f697d5a5bb01

The Nexus remote will disconnect during encryption when paired with the
secure connections feature enabled. This patch will blacklist the remote
and mask out the feature during pairing.

This patch introduces a new interop database with the aim of
consolodating various blacklists and interop workarounds into a single
database for code readability and maintainability of the list.

Bug: 21817410
Change-Id: I6f510e1c8c6050ab4e313209122d028e12e5f54f

This patch uses the transport type from the GATT event instead
of using the device type stored in NVRAM, to determine the
transport type while encrypting the link. The issue is seen
with dual mode devices, when the connection on LE transport
causes encryption request on the BR/EDR transport

Bug: 21652097
Change-Id: I895756f9c0048b5a559ea00f633b1100109e52b8

This change prevents the LE RNR at the end of inquiry. This prevents
extra LE connections created during BT scan, to fetch remote name,
which will be performed later during connections.

Bug: 22228981
Change-Id: Ied61bc192cec611e7b2f41e474c5f032ed9f0981

Change Own address type to Random identity only if controller supports enhanced
privacy. Check if controller supports enhanced privacy before changing own adress type
and peer address type to 0x02 or 0x03 in auto connection.
Privacy 4.1 only supports Public and Random address.

Bug: 20817308
Change-Id: I9f5531e4e41c824426cd64de517b44355ac2655d

The  entries on the "command_queue" are allocated by osi_calloc()
hence they should be deallocated by osi_free()

Bug: 21784321
Change-Id: Ic6e13dbafef2c6ac79ce74ebfc4db702c9ef04c5

Bug: 21756298
Change-Id: I4bd38f6fa594ccb4dd9f23f3887484743a3bc451

Don't reuse AVRC buffers for responding back. Apparently,
in the AVRC packets we transmit the offset with the payload must
be at least 15 octets. If the original buffer is not large enough,
this results in memory corruption.

Also, use an explicit check for 'bt_rc_ctrl_callbacks' as a workaround
until the upper layer does the right thing with the callbacks registration.

Bug: 22006014

Change-Id: I28c248d1580bdddbda76298d19faadf8985187fc

This works around a race condition in which the just-connected remote
device sends the local device an indication before the appropriate
handle/device/etc. has been added to the cache.  Previously we were
dropping that indication, which led to the remote device timing out
and disconnecting some time after the connection had been successfully
established.

Bug: 21026847

Change-Id: Iea43e7c93e48b5e7a7e78a1c3fb591d6fe972fc3

Use Case:
1) pair and Connect to HOGP mouse
2) Disconnect from setting menu.

Failure:
Disconnect is not sent immediately

Root Cause:
While disconnecting we start lcb's default timeout of 4 seconds before
disconnecting the link. Since we are removing fixed channels associated
to a remote, setting idle timeout to 0 for immediate disconnection.

Bug: 21877809
Change-Id: I39fedb6cac6f952149cb2722d028115fbdc20b70

Issue:
-> In AV start-up sequence, State machine initialization was being
called after enabling Btif A2dp service.
-> In that case, State machine handle will be empty and it will
prevent enable and register events to enter into BTIF.
-> Because of the absence of enable/register events, AV module in
stack will be unaware of upper layer registration and will not
proceed for connection.

Fix:
-> Changed the sequence in AV start-up, initialized State machine
first before dispatching any events to State machine.
-> So, the handle will not be empty and will not block events like
BTA_AV_ENABLE_EVT and BTA_AV_REGISTER_EVT to enter into BTIF.

Bug: 21451237
Change-Id: I5d47e20ed68ea5730a7c149c316446eb8a1d41af

Prevents GKI Exception in BT stack when phone enters suspended state
after performing HOGP device connection by clearing the allocated
memory when write to an unmatched report is done.

Bug: 22028876
Change-Id: Ia6ca9405142410e894c7ebe84d8827ffdc7659ba

The SMP slave state machine goes on to generate local nonce
after a successful commit, and ignores any commit from the
remote master if received during the generation process.
This change fixes that race condition

Bug: 21896912
Change-Id: I0cc2f1d20f6754d19f8b39fd62e3176007c50f1b

This patch ensures that Bluetooth Inquiry uses the low latency
scan parameters, and the scan params set by the Apps for LE scan
are retained and set at the end of a BT inquiry.

Bug: 21896912
Change-Id: Iaf9e47330393cc1e0c21c625e8e07ce933506983

Enable address resolution to resolve the incoming connection details
to provide the correct address type as part of the LE connection
event.

Bug: 21751039
Change-Id: I148b0ffbfa1a39a32d90f5607362754d1901203f

Use case: Test the secure connections only mode PTS test cases.

Test Steps:
1. Call the API BTM_SetSecureConnectionsOnly from BT-IF layer to
   put the DUT is under secure connections only mode to execute
   the PTS test case TP/SEC/SEM/BV-11-C.
2. Run the required steps for PTS test case.

Failure: Failed to enter into secure connections only mode.

Root Cause: The function BTM_SetSecureConnectionsOnly doesn't set
btm layer security mode variable to only mode and all further checks
are based on that variable.

Fix: Fixed it by setting the security mode variable properly
in the API function.

Bug: 21896912
Change-Id: I21d37e7d24c443c2cbcd72d6f62f5e31fa7cf107

Use case:   Check the A2dp play , pause from headset

STR:
Connect to the headset from DUT
Try play , pause from Headset once the song is started playing from headset.

Failure: Bluetooth process crashed due to invalid fd descriptor while
clearing fds using FD_CLR

Root cause: Root cause for this issue is pthread join mechanism for
UIPC thread is not proper ( incorrect logic ), as a result UIPC thread
still running, and A2DP media task will try to start a new UIPC thread
before the previous UIPC read thread is closed, finally sometimes
this scenario leads to this issue. This issue doesn't come always,
if the media task is in process of initiating the fds before staring
the new UIPC thread, and previous UIPC thread is in exiting state.

Fix: Correct the logic of UIPC thread id check while joining the UIPC thread.
Thread id might hold pointer value where it's value is negative vaule with
singed bit is set,so corrected the logic to check against zero or  non zero.

Bug: 21896912
Change-Id: I1307d848958656e718e95a972f258526470b1974

Use case: Test the paring scenario with remote device with security mode 3.

STR:
1. Create a connection from Remove dev with Security Mode-3 to DUT.
2. Enter pin key in remote, when pin key pop-up comes in DUT, reject the same.
3. Try to create connection from DUT to Remote dev-2, within 30 seconds.

Failure: Create connection not sent from DUT to remote, even though
there was no pairing taking place

Root cause: As the pairing state is not reset properly it disturbs
the next pairing process.

Reproducible using MecApp test application.

Fix: Reset the pairing state to idle in case stored paring
device matches with the device for which the pairing
failure happened.

Bug: 21896912
Change-Id: I0a9e143a44bb2ae7e296d7d76c1323e517f9cc42

Use case: Test the connect collision scenario from DUT and Headset.

STR:
1. Create a connection from DUT to headset, disconnect and unpair the link.
2. Turn OFF and turn ON headset, keep in pairing mode, headset keeps trying
   to connect back to DUT.
3. Inquiry in DUT, and connect to headset.

Failure: DUT gets stuck in pairing, whenever there is a clash in
accept_connection and create_connection.

Root cause: When DUT and remote try to connect to each other simultaneously,
an ACL handle is created with first connect request and the second connect
request is also processed.DUT doesn't check whether there is already a
ACL handle present between the devices.

Fix: Fixed the issue by reusing the existing handle and
discarding the second connect request.

Bug: 21896912
Change-Id: I2001bc3b2fa50f827a4e3b41cde91af1ea106c54

Use case: Allocated required GKI buf in sdp attr req

1. Enter UUID - 0100 for L2CAP or 110C for AVRCP Target so DUT sends
   continuation frame for TSPX_sdp_service_search_pattern on PTS
2. Start Test case TP/SSA/BV-06 in PTSv6.0

Failure: Some PTS test cases request less attributes in first packet
and request more attributes in continuation packets.
As stack allocates the buf in start packet and using the same
buf in continuation packets, it's causing buffer corruption and crash

Root cause: Buffer allocated for start packet is not sufficient
in continuation packets

Fix: Fixing this issue by dynamically allocating buffer in continuation
packets of service_search_attr_req and service_attr_req

Bug: 21896912
Change-Id: I8daeffb7d6486c7b916ad2f0505ad422d91a613c

Use case: NULL pointer check in sdpu_build_attrib_entry

Crash observed during BT stability test

Failure: Crash observed while accessing p_attr->value_ptr

Root cause: There is no null check for p_attr->value_ptr while
accessing it in sdpu_build_attrib_entry

Fix: Fixing this issue by adding null check for p_attr->value_ptr
while accessing it in sdpu_build_attrib_entry

Bug: 21896912
Change-Id: I1dd352a33ff6f86add7d1b1bfdf240d1b7992b83

Host can send connection request when paired HID device
is in disconnected state. We are not checking the
NORMALLY_CONNECTABLE flags from sdp record, and by default
sending this request from host, for subsequent user initiated
connection to paired remote HID Devices.
If the remote HID is in page scan mode, it will get connected
and if not, connection will fail after retrying until
HID_HOST_MAX_CONN_RETRY times.

Without this patch, if remote hid device sdp record shows that
device's NORMALLY_CONNECTABLE flag is false, host will not be
able to create outgoing connection to remote device even if the
device is in pairing mode until device is unpaired giving bad
user experience.

Bug: 22028876
Change-Id: I2b3c5c25dc7b08ab6ed0c3667897e5ea0f05f914

A case, where 3 remote device (PANU role) is connected with DUT (NAP role).
the one remote (PANU role) send BNEP_COMPRESSED_ETHERNET_DEST_ONLY
packet to DUT (NAP role), which is for another remote device (PANU role).
So DUT (NAP role) forwards this packet to another remote device (PANU role).
As same pbuf is used across the L2CAP/BNEP/PAN layer since reception from
L2CAP to writing on L2CAP CID of another PANU. Sometimes buffer (p_buf) gets
corrupted, when required buffer size is greater than actual buffer size,
while rebuilding the Bnep header.

Using bnep_write instead of bnep_writeBuf to solve the buffer corruption
issue, which allocates the new buffer & append data into it.

This patch also add TRC_BNEP & TRC_PAN into conf file.

Bug: 22028876
Change-Id: I125cab6d3da0f8126605bc7e9f764a27ee72fb1f